The proposed research will examine the relationship between the social environment and frequency of drug use among a predominantly African-American sample of adult active drug users. Because drug use/abuse is associated with a number of negative health and social consequences-increased crime rates in communities, unemployment, mental health disorders and HIV disease--a greater understanding of the predictors of drug use frequency is essential for the development of successful secondary prevention strategies. The proposed research will employ multi-level techniques to determine the individual, network, and neighborhood level factors that may influence variations in frequency of use. Theories of community social disorganization and social influence processes will provide a basis from which to address the following questions: Do neighborhood characteristics have a direct association with drug use frequency? How do characteristics of drug networks modify the relationship between neighborhood characteristics and frequency of use? How do the multiple levels of influence interact to affect frequency of use? Addressing these questions will fill gaps in existing literature and contribute to current knowledge of the impact of social context on drug use. Data to be used are from two sources: 1) SHIELD, an HIV intervention project in Baltimore, MD and publicly available data. Investigation into relevant socio-environmental contexts (peer networks, neighborhood), in addition to individual predictors, can inform policy decisions and action plans for effective interventions.